For the last 30 years, the actuaries who draw up the long-run projections for Social Security have been forecasting that by the 2030s, there were be inadequate funds to pay promised Social Security benefits. There is no secret here–but nothing has been been done. Douglas Arnold describes the situation and makes some judicious predictions about what is likely to happen in “Fixing Social Security: The Politics of Reform in a Polarized Age,” (Milken Institute Review, Third Quarter 2022). It’s an excerpt from Arnold’s just-published book of the same title. Arnold writes:

We should not be surprised if Congress does nothing to fix Social Security before 2034, when the trust fund runs dry. Although experts first identified the long-term solvency problem nearly three decades ago, and opinion surveys have repeatedly shown that fixing the problem is one of the public’s top priorities, legislators have never voted on a proposal to fix it — not in committee, not on the floor, not in the House, not in the Senate. … And the principal reason for congressional inaction is clear: insolvency is a long-term problem without short-term consequences. Everything will change in 2034. Suddenly, insolvency will become an urgent problem with enormous consequences. Absent congressional action, an estimated 83 million Social Security recipients — 18 million more than today — will face automatic benefit cuts of 21 percent. Another 8 million people filing for Social Security benefits that year will face similar reductions from what they would otherwise collect.

It’s worth emphasizing that Arnold’s comment that “legislators have never voted on a proposal to fix it.” Apparently, both parties would prefer to have the Social Security issue linger, rather than find a way to take credit for fixing it. This seems especially striking to me because addressing the coming shortfall in Social Security is pretty straightforward. The same actuaries who point out that the system isn’t financially sustainable as it stands also offer financial estimates of a list of possible policy choices. I’ve written about these kinds of proposals before (for example, see here and here) and won’t run through them again here. But I sometimes say that if a bipartisan group was locked in a room one morning and told that they wouldn’t be served food until a compromise was reached, I think the group could easily be out with a plan before lunchtime.

If some mixture of proposals like these for a later retirement age, or an increase in the amount of income on which the payroll tax is levied had been adopted a 5 or 10 or 15 years ago, then as the effect of modest changes accumulated over time, Social Security could have been made solvent for many decades into the future with hardly anyone noticing. As the 2034 timeline approaches, the necessary changes have become bigger and bigger.

Arnold argues that Congress is unlikely to take action before the Social Security funding crisis is upon us. After all, the previous time that the Social Security trust fund was about to run out of money, in the early 1980s, Congress waited until the last minute and then appointed a commission (under the leadership of Alan Greenspan, who later became chair of the Fed) to propose a solution. Arnold points out that there are special rules in the federal budget process which require that any changes to Social Security will need to get 60 votes in the US Senate–that is, the changes cannot be made by a simple majority as part of the budget process. Thus, both parties will likely need to sign off.

When Congress decided to show its bravery by appointing another commission in about 2034, what choices will at that point be on the table?

There is a subgroup in both parties that would like to make relatively substantive changes to Social Security. On the Republican side, there is a group that is eager to transform much or all of Social Security into a set of individual retirement accounts, where the federal government would top up the accounts for those with low incomes. For example, if a proposal along these lines had been implemented about 15 years ago, so that holders of these retirement accounts could have benefited from the long run-up in the stock market since about 2010, a lot of people would be feeling a lot better about their retirements just now. But individual accounts would also create a need for a snake’s nest of rules about how such accounts could be invested, if one could use them as collateral for loans, if people would be allowed to dip into them for “worthy” purposes like house down-payments or college tuition for their children or paying legal settlements–and so on and so on.

Most Democrats are resolutely against altering Social Security in this way, but a certain subset of Democrats would like to see the benefits of the system substantially expanded. Because Social Security payments are linked to the taxes a person (or a spouse) paid into a system during a working, those who didn’t pay much into the system can end up in deep poverty when they are older. Of course, when a system is already on a track for a financial crash, a substantial addition to the benefits it would pay out would make the financial crunch worse.

By 2034, if nothing happens until then, Social Security beneficiaries would be facing a permanent drop of about 21% in their benefit checks. The gap to make up the difference would be about 1.2% of GDP for every year into the future. What is likely to happen?

Well, it would presumably be political suicide for politicians if Social Security benefit rates declined. Thus, while one can imagine longer-term changes in benefits, like a slow phase-in of a later retirement age, or changes in the details of how benefits are calculated. Over a few decades, these can make a big difference. But in the moment of the crisis in 2034 it’s unlikely that current benefits will be cut in any meaningful way.

On the tax side, a number of current Republicans have staked out ground that they will not support an increase in payroll taxes. Again, one can imagine policies that might have the effect of a slow phase-in of higher taxes–say, increasing the income taxes that those with high incomes might pay on Social Security benefit–but in the moment of crisis in 2034, a jagged upward jump in taxes for the system also seems unlikely.

Back when the Social Security system was being saved in the 1980s, one short-term step was to shuffle some money around from other federal trust funds for a few years. But the other federal trust funds are not flush with money, either. The Medicare trust fund is scheduled to run out in 2028. Borrowing from the federal trust funds for retired civilian and military personnel is not a long-run fix.

Thus, a plausible prediction for 2034 is that Social Security will be “fixed” by turning to general fund tax revenues–rather than the payroll tax–as a source of funding. I suspect this would be done with a lot of strong statements about how it was only a temporary change, but it’s the kind of temporary change that can easily become permanent. As Arnold points out, this outcome is plausible–and would also represent a major change to the operation of Social Security:

Policymakers have had good reasons for not using general funds to subsidize Social Security. President [Franklin]Roosevelt argued that a tight link between taxes and benefits served two important ends. It would protect Social Security from hostile actors — “No damn politician can ever scrap my Social Security program” — but it would also protect the program from unreasonable expansion. Legislators could not expand the program unless they were also willing to increase taxes.

This tight link has worked for nearly a century. The program’s detractors have never found a way to dismantle Social Security because workers earn their benefits by paying a dedicated tax. But neither have the program’s champions been able to expand benefits since 1972 because legislators have been unwilling to increase taxes.

As someone whose current life plan is to start drawing Social Security benefits in 2030, when I turn 70, I contemplate these possibilities with little cheer.